#604932 segfaults when encoding to vorbis on armel

#604932#5
Date:
2010-11-25 14:16:41 UTC
From:
To:
Hi,
When connecting with a remote client to mpd it crashes on when playing the first song like:

(gdb) thread apply all bt
Thread 4 (Thread 0x43a2e410 (LWP 3082)):
#0  0x40e13d0c in vorbis_encode_setup_init () from /usr/lib/libvorbisenc.so.2
#1  0x40e14088 in vorbis_encode_init () from /usr/lib/libvorbisenc.so.2
#2  0x0001da2c in ?? ()
#3  0x0001da2c in ?? ()
Backtrace stopped: previous frame identical to this frame (corrupt stack?)

Thread 3 (Thread 0x4302a410 (LWP 3081)):
#0  0x00017fb0 in ?? ()
#1  0x00000400 in ?? ()
#2  0x00000400 in ?? ()
Backtrace stopped: previous frame identical to this frame (corrupt stack?)

Thread 2 (Thread 0x4282a410 (LWP 3080)):
#0  0x410e05f0 in pthread_cond_wait@@GLIBC_2.4 () from /lib/libpthread.so.0
#1  0x00026978 in ?? ()
Cannot access memory at address 0x2
#2  0x00026978 in ?? ()
Cannot access memory at address 0x2
Backtrace stopped: previous frame identical to this frame (corrupt stack?)

Thread 1 (Thread 0x41e07b60 (LWP 3077)):
#0  0x41490f68 in poll () from /lib/libc.so.6
#1  0x402a594c in g_poll () from /lib/libglib-2.0.so.0
#2  0x40298600 in ?? () from /lib/libglib-2.0.so.0
#3  0x40298600 in ?? () from /lib/libglib-2.0.so.0
Backtrace stopped: previous frame identical to this frame (corrupt stack?)

The backtrace isn't very helpful though. I'd be happy to provide more
information if necessary. Note that the system in question is an arm box. It
doesn't make a difference if I use a local or a remote client. Browsing the
songs in the library works fine.

Filing this as grave since it looks completely unusable on this arch.

Cheers,
 -- Guido

#604932#10
Date:
2010-12-01 10:58:33 UTC
From:
To:
Just to add my 2 cents to that bug. I have the exact same bug, with
the same arch.

When I mean the same bug, I have the same conf and the same version.
The core files is the same.

It seems that libvorbisenc on armel is not working...

Cheers everyone

#604932#15
Date:
2010-12-12 18:06:08 UTC
From:
To:
severity 604932 important
thanks

This bug is not RC -- mpd works ok on arm with the default configuration,
it's only re-encoding as a streaming server that suffers. I suspect many
arm systems are too slow to be able to encode to vorbis anyway, even
with an integer encoder. So this seems a niche problem.

#604932#22
Date:
2012-11-26 23:46:04 UTC
From:
To:
Hi Maxime and Guido,

Maxime Hadjinlian wrote (01 Dec 2010 10:58:33 GMT) :

Can any of you reproduce this bug on current testing/sid?

Cheers,
--
  intrigeri
  | GnuPG key @ https://gaffer.ptitcanardnoir.org/intrigeri/intrigeri.asc
  | OTR fingerprint @ https://gaffer.ptitcanardnoir.org/intrigeri/otr.asc

#604932#27
Date:
2012-11-27 11:40:58 UTC
From:
To:
Hi
Yep, tried just a moments ago and still the same problem.

#604932#32
Date:
2012-11-27 12:14:45 UTC
From:
To:
Maxime Hadjinlian wrote (27 Nov 2012 11:40:58 GMT) :

May you please tell us with what version of the mpd package, exactly,
you just reproduced this, so that I mark the bug as found in
that version?

#604932#37
Date:
2012-11-27 12:33:04 UTC
From:
To:
With :
    - mpd (MPD: Music Player Daemon) 0.16.7
    - the package version is 0.16.7-2 (armel)

#604932#52
Date:
2024-08-26 05:39:02 UTC
From:
To:
Are anyone able to reproduce this issue using valgrind?  Perhaps it
can provide more details on the crash?  Is there a way to reproduce it
without a full mpd setup?